Real Urban Environment Research Articles

Abstract Today, machine learning has a crucial role in wireless communications, notably in 5G and 6G. It contributes significantly for increasing network capacity, improving user experience, and enhancing network reliability. Among machine learning techniques, reinforcement learning is vital due to its suitability for many real-world scenarios. It enables agents to learn from the environment with zero-knowledge and make rational decisions. Thus, in this article, we aim to explore the role of classical reinforcement learning in predicting optimal beam angles within urban environments. The goal is to minimize interference between antennas by finding optimal beamforming angles using ray tracing techniques. We examine various classic reinforcement learning methods in an urban scenario, focusing on maximizing total channel capacity. Initially, we identify the optimal beamforming angles for maximizing channel capacity with four antennas. After validating the learning methods and achieving over 99% accuracy, we proceeded to utilize them in a larger scenario. In the first phase, these methods and their accuracy are validated based on the results of the exhaustive search for a small number of nodes. In the second phase, we predict optimal antenna beam angles for scenarios with an increased number of transmitters and receivers for a realistic urban environment situated in the north-eastern part of Berlin.

A deep-learning-based closure model to address energy loss in low-dimensional surrogate models based on proper-orthogonal-decomposition (POD) modes is introduced. Using a transformer-encoder block with an easy-attention mechanism, the model predicts the spatial probability density function of fluctuations not captured by the truncated POD modes. The methodology is demonstrated on the wake of the Windsor body at yaw angles of $\delta = [2.5^\circ ,5^\circ ,7.5^\circ ,10^\circ ,12.5^\circ ]$ , with $\delta = 7.5^\circ$ as a test case, and in a realistic urban environment at wind directions of $\delta = [-45^\circ ,-22.5^\circ ,0^\circ ,22.5^\circ ,45^\circ ]$ , with $\delta = 0^\circ$ as a test case. Key coherent modes are identified by clustering them based on dominant frequency dynamics using Hotelling’s $T^2$ on the spectral properties of temporal coefficients. These coherent modes account for nearly $60 \,\%$ and $75 \,\%$ of the total energy for the Windsor body and the urban environment, respectively. For each case, a common POD basis is created by concatenating coherent modes from training angles and orthonormalising the set without losing information. Transformers with different size on the attention layer, (64, 128 and 256), are trained to model the missing fluctuations in the Windsor body case. Larger attention sizes always improve predictions for the training set, but the transformer with an attention layer of size 256 slightly overshoots the fluctuation predictions in the Windsor body test set because they have lower intensity than in the training cases. A single transformer with an attention size of 256 is trained for the urban flow. In both cases, adding the predicted fluctuations close the energy gap between the reconstruction and the original flow field, improving predictions for energy, root-mean-square velocity fluctuations and instantaneous flow fields. For instance, in the Windsor body case, the deepest architecture reduces the mean energy error from $37 \,\%$ to $12 \,\%$ and decreases the Kullback–Leibler divergence of velocity distributions from ${\mathcal{D}}_{\mathcal{KL}}=0.2$ to below ${\mathcal{D}}_{\mathcal{KL}}=0.026$ .

ABSTRACTThe use of underground freight transportation (UFT) is gaining attention because of its ability to quickly move freight to locations in urban areas while reducing road traffic and the need for delivery drivers. Since packages are transported through the tunnels by electric motors, the use of tunnels is also environmentally friendly. We examine the use of tunnels to transport individual orders, motivated by the last‐mile delivery of goods from e‐commerce providers. The use of UFT for last‐mile delivery requires more complex network planning than for direct lines that have previously been considered for networks connecting large cities. We introduce a new network design problem based on this delivery model and transform the problem into a fixed‐charge multicommodity flow problem with additional constraints. We show that this problem, the ‐UFT, is NP‐hard, and provide an exact solution method for solving large‐scale instances. Our Benders decomposition‐based solution approach exploits the combinatorial sub‐structures of the problem in a cutting planes fashion. We provide computational results for real urban environments to build a set of insights into the structure of such networks and evaluate the benefits of such systems. We estimate the costs for implementing UFT systems and break them down into a per‐package cost. Our estimates indicate at least a 40% cost savings from using a UFT over traditional delivery models. This indicates that UFT systems for last‐mile delivery are a promising area for future research.

Traffic assignment in large urban areas is an old but increasingly important problem because of the rapid growth of the world population and traffic demands. Many algorithms have been developed but their convergence rates and complexities are still prohibitive for real-time applications. The recently developed MinSum algorithm introduces a new approach. It is a highly efficient discrete-domain optimization algorithm for system-optimized route assignment between two city zones. Its complexity (the number of critical operations) is O(R3), where R is the number of routes. Nonetheless, there is still room for improvements, and this paper presents two modified MinSum variants, heuristic and approximate, that are significantly faster and of lower complexity, while retaining MinSum’s prominent features. Heuristic variant MinSumH is up to five times faster than MinSum and its complexity theoretically remains O(R3), though experiments indicate that it is closer to O(R2). Approximate variant MinSumA is faster by up to over 100 times and reduces the complexity to O(R). Both proposed variants are progressively faster as R grows. Due to their high convergence rate and exceptionally low complexity, along with other prominent features, the proposed algorithms are ready for real-time system-optimal traffic assignment in a real urban environment.

The Internet of Things (IoT) has enabled a vast network of devices to communicate over the Internet. However, the fragmentation of IoT systems continues to hinder seamless data sharing and coordinated management across platforms.However, there is currently no actual search engine for IoT data. Existing IoT search engines are considered device discovery tools, providing only metadata about devices rather than enabling access to IoT application data. While efforts such as IoTCrawler have striven to support IoT application data, they have largely failed due to the fragmentation of IoT systems and the heterogeneity of IoT data.To address this, we recently introduced SensorsConnect-a unified framework designed to facilitate interoperable content and sensor data sharing among collaborative IoT systems, inspired by how the World Wide Web (WWW) enabled shared and accessible information spaces for humans. This paper presents the IoT Agentic Search Engine (IoTASE), a real-time semantic search engine tailored specifically for IoT environments. IoTASE leverages LLMs and Retrieval-Augmented Generation (RAG) techniques to address the challenges of navigating and searching vast, heterogeneous streams of real-time IoT data. This approach enables the system to process complex natural language queries and return accurate, contextually relevant results in real time. To evaluate its effectiveness, we implemented a hypothetical deployment in the Toronto region, simulating a realistic urban environment using a dataset composed of 500 services and over 37,000 IoT-like data entries. Our evaluation shows that IoT-ASE achieved 92% accuracy in retrieving intent-aligned services and consistently generated concise, relevant, and preference-aware responses, outperforming generalized outputs produced by systems such as Gemini. These results underscore the potential of IoT-ASE to make real-time IoT data both accessible and actionable, supporting intelligent decision-making across diverse application domains.

In urban areas characterized by complex subsurface conditions, such as Agargaon in Dhaka, Bangladesh, efficient and accurate simulation of porous media is crucial for effective urban planning and management. This article introduces a refined numerical simulation method to address the challenges of modelling porous media in real urban environments. The proposed method combines advanced grid techniques with porous and standard procedures for solving flow and transport equations in porous media. By integrating high‐resolution spatial data with a multi‐level approach that optimizes computational resources, the method achieves significant reductions in simulation time while preserving accuracy. Applying this method to Agargaon demonstrates its effectiveness in modelling substantial improvements in predicting pollutant dispersion and urban heat island effects and predicting wind flow, pollutant dispersion and urban heat island effects within the urban context. This approach provides a practical solution for urban planners and engineers, offering enhanced analytical capabilities to support sustainable development in rapidly growing cities.

Urban building energy prediction is a critical challenge for sustainable city planning and large-scale retrofit prioritization. However, traditional data-driven models struggle to capture real urban environments’ spatial and morphological complexity. In this study, we systematically benchmark a range of graph-based neural networks (GNNs)—including graph convolutional network (GCN), GraphSAGE, and several physics-informed graph attention network (GAT) variants—against conventional artificial neural network (ANN) baselines, using both shape coefficient and energy use intensity (EUI) stratification across three distinct residential districts. Extensive ablation and cross-district generalization experiments reveal that models explicitly incorporating interpretable physical edge features, such as inter-building distance and angular relation, achieve significantly improved prediction accuracy and robustness over standard approaches. Among all models, GraphSAGE demonstrates the best overall performance and generalization capability. At the same time, the effectiveness of specific GAT edge features is found to be district-dependent, reflecting variations in local morphology and spatial logic. Furthermore, explainability analysis shows that the integration of domain-relevant spatial features enhances model interpretability and provides actionable insight for urban retrofit and policy intervention. The results highlight the value of physics-informed GNNs (PINN) as a scalable, transferable, and transparent tool for urban energy modeling, supporting evidence-based decision making in the context of aging residential building upgrades and sustainable urban transformation.

Vehicular ad hoc networks (VANET) are revolutionizing transportation by enabling real-time communication between vehicles and roadside infrastructure, enhancing safety and efficiency through the exchange of traffic updates, road conditions, and critical data. However, VANET faces significant security threats, including Sybil, black hole, and wormhole attacks, where malicious nodes manipulate network communication, leading to misinformation and disruptions. This research proposes a novel hybrid model, GBiL, integrating gated recurrent unit (GRU) and bidirectional long short-term memory (BiLSTM) to detect and mitigate such attacks. At the core of this architecture, an intrusion detection system (IDS) is combined with a trust detection module to assess the trustworthiness of network nodes using real-time data. The IDS employs a hybrid approach for efficient intrusion detection, leveraging particle swarm optimization (PSO) after feature selection. To ensure a balanced dataset, data augmentation is applied using SMOTETomek, a combination of Synthetic Minority Over-sampling Technique (SMOTE) and Tomek Links. Real-time simulations using NS-3 and SUMO with real-world mapping from OpenStreetMap validate the system’s effectiveness in a realistic urban network environment by generating a dataset called VANET Attacks (VA) dataset. This comprehensive approach strengthens VANET security against multiple attack vectors. The proposed GBiL model achieves high performance, with an accuracy of 96.01% and a false alarm rate of just 0.04%. This research significantly enhances VANET security by integrating sophisticated detection techniques, data augmentation, and real-time trust evaluations, establishing a robust foundation for more secure and reliable autonomous transportation through improved vehicular communication networks.

Differences between time-averaged and ensemble-averaged wind are studied for the case of changing wind direction. We consider a flow driven by a temporally turning pressure gradient in both an idealized case of a staggered cube array and a realistic urban environment. The repeating structure of the idealized case allows us to construct a large ensemble of 3240 members with a reasonable compute time. The results indicate that the use of plain time averaging instead of an ensemble average can severely reduce the accuracy of both the mean and variance. These errors are the largest when the averaging time is of the same order as the time scale associated with the turning. Utilizing Taylor diagrams, we show that a reasonable compromise between ensemble size and accuracy can be achieved by calculating the ensemble statistics from temporally averaged results with an averaging time that is clearly smaller than the characteristic time scale. This allows the use of reasonably-sized ensembles with 10–50 members. By applying this approach to the realistic urban geometry, we identify building wakes as the regions most severely affected by the incorrectly use of time averaging.

Numerical investigation of pollutants distribution in a realistic urban environment: A Frankfurt city study

ABSTRACTIn smart cities, a common problem is the parking spots classificationinto empty and occupied. It may seem simple, but a large numberof Deep Learning approaches rely on CNNs (Convolutional NeuralNetworks). These solutions are commonly expensive, demandinghigh computational power and specialized hardware to run properly,making them unsuitable for large-scale deployments, such asin smart cities. In this work, we propose two lightweight CNN architectures,built upon existing solutions by enhancing their efficiencyand robustness. We used a cross-dataset scenario, where a modelis trained and validated in two datasets and tested in another, applyingthree robust state-of-the-art datasets: PKLot, CNRPark-EXTand PLds. This process improves generalization across differentcontexts and sets a more realistic scenario when compared to realurban environments. Also, we compared our models to state-ofthe-art networks, such as MobileNetV3 Large and Small to ensureconsistency and validate the results with well-explored modelsin the literature. Our results showed that our models, with up to34× and 88× fewer parameters than the MobileNetV3 Large, reachless than 2% lower accuracy when compared to the MobileNetV3networks. Furthermore, by using grayscale images, the results wereslightly better and also decreased processing and storage costs.

Urban distributed energy systems play a crucial role in the development of sustainable and low-carbon cities. Evaluating urban wind resources is essential for effective wind energy harvesting, which requires detailed information about the urban flow field. Computational fluid dynamics (CFD) has emerged as a viable and scalable method for assessing urban wind resources. This review paper synthesizes the characteristics of the urban wind environment and resources, outlines the general framework for CFD-aided wind resource assessment, and addresses future challenges and perspectives. It highlights the critical need to optimize wind energy harvesting in complex built environments. The paper discusses the conditions for urban wind resource assessment, particularly the extraction of boundary conditions and the performance of small wind turbines (SWTs). Additionally, it notes that while large eddy simulation (LES) is a high-fidelity model, it is still less commonly used compared to Reynolds-averaged Navier–Stokes (RANS) models. Several challenges remain, including the broader adoption of high-fidelity LES models, the integration of wake models and extreme conditions, and the application of these methods at larger scales in real urban environments. The potential of multi-scale modeling approaches to enhance the feasibility and scalability of these methods is also emphasized. The findings are intended to promote the utilization and further development of CFD methods to accelerate the creation of resilient and energy-efficient cities, as well as to foster interdisciplinary innovation in wind energy systems.

Urban densification associated with rapid urbanization has weakened horizontal ventilation in cities. Previous studies point out that building-height variability can enhance vertical ventilation, while most of them rely on idealized models that overlook the complexity of real urban environments. This study analyzes 20 actual urban blocks using CFD simulations, considering average building height, building density, and height standard deviation. The results show that areas with low-rise, uniform buildings exhibit superior pollutant dispersion, while mid- and high-rise zones experience complex turbulence and pollutant accumulation. Ventilation performance peaks when the height standard deviation ranges between 35 and 40. These findings underscore that optimizing urban form for vertical ventilation requires a combined strategy of density control and height variation. Realistic building group models more accurately capture airflow dynamics and provide valuable insights for the design of effective vertical ventilation corridors and the enhancement of urban pollutant dispersion.

Given the ongoing trend of urbanization and the increased prevalence of specific mental disorders in urban settings, there is a need to better understand the link between urban living and mental health. Recent advances in urban mental health research have leveraged mobile electroencephalography to explore how brain electrical signals are influenced by urban stressors and resources. This study aims to synthesize the evidence from mobile electroencephalography measurements in the context of urban mental health. A systematic literature research was conducted in the databases PubMed/MEDLINE, Embase, PsycINFO and CINAHL in September 2023. The present review includes primary studies that used in-situ electroencephalography in real urban environments published since 2013. Four independent reviewers conducted the screening, while two researchers performed data extraction using Microsoft Excel and assessed risk of bias using the Effective Public Healthcare Panacea Project Quality Assessment Tool. The review has been pre-registered with the International Prospective Register of Systematic Review (PROSPERO) under the registration number CRD42023471636. Fifteen studies were identified, primarily examining power in alpha, beta, and theta frequencies in urban areas compared to less urbanized environments. Study findings exhibited significant heterogeneity; while some studies noted heightened brain activity in urban environments, others observed reductions compared to less urbanized or greener regions. Notably, certain demographic cohorts, such as adolescents, have been understudied. Moreover, descriptions of exposures were often inadequate for ensuring replicability, and gender considerations were seldom integrated into analyses. This systematic review provides insights into an emerging field of research which appears to be suffering from small sample sizes and a lack of methodological transparency and consistency. Interpretation of the seemingly contradictory results requires future studies to be more rigorous in documenting urban exposures and choice of brain components under investigation.

The automotive industry is constantly evolving, demanding advanced technologies for growth. Among these is cellular technology's long-term evolution (LTE), which 3GPP is pioneering to provide vehicle-to-everything (V2X) connectivity. It concentrates on vehicle-to-vehicle (V2V) direct communications, essential for safety applications like traffic management or infotainment. Something significant here is having a system that allows them to understand each other and cooperate. All vehicles should be able to do this by sending messages about what they are doing and where they are so that it will be safer for drivers moving through town where there may be several hurdles that break the signal lines or block them out completely. Still, before this goal can be reached, there is a need for clever ways of using packets to maximize radio channels while ensuring the service works correctly. This task poses a considerable challenge while constructing algorithms that can handle it, which is an open development question. To solve this development question, a simulator named LTEV2Vsim has been introduced. Written in MATLAB, this simulator uses straightforward models for managing vehicle movement or more complex ones based on realistic urban environment input files. This simulator is helpful for scholars interested in validating and designing LTE-V2V networks. The ability to simulate urban environment scenarios and give example results will help develop a better system for carrying packets and thus ensure that cooperative awareness services are successfully implemented in V2X communication.

Abstract. High-rise buildings, increasingly a feature of many large cities, impact local atmospheric flow conditions. Tall-building wakes affect air quality downstream due to turbulent mixing and require parameterisation in dispersion models. Previous studies using numerical or physical modelling have been performed under idealised and neutral conditions. There has been a lack of data available in real urban environments due to the difficulty in deploying traditional wind sensors. Doppler wind lidars (DWLs) have been used frequently for studying wind turbine wakes but never building wakes. This study presents a year-long deployment of a DWL in a complex urban environment, studying tall-building wakes under atmospheric conditions. A HALO Photonics StreamLine DWL was deployed in a low- and mid-rise densely packed area in central London. From its rooftop position (33.5 m a.g.l. compared to mean building height of 12.5 m), velocity azimuth display (VAD) scans at 0° elevation intersected with two taller nearby buildings of 90 and 40 m a.g.l. Using an ensemble-averaging approach, wake dimensions were investigated in terms of wind direction, stability, and wind speed. Boundary layer stability categories were defined using eddy covariance observations from the BT Tower (191 m), and mixing height estimations were made from vertical stare scans. A method for calculating normalised velocity deficit from VAD scans is presented. For neutral conditions, wake dimensions around both buildings for the prevailing wind direction were compared with the ADMS-Build wake model for a single, isolated cube. The model underpredicts wake dimensions, confirming previous wind tunnel findings for the same area. Under varying stability, unstable and deep boundary layers were shown to produce shorter, narrower wakes. Typically observed wake lengths were 120–300 m, and widths were 80–150 m and reduced by 50–100 m downwind. Stable and shallow boundary layers were less frequent and produced an insignificant difference in wake dimensions compared to neutral conditions. The sensitivity to stability was weakened by enhanced turbulence upstream (i.e. due to other building wakes). Weakened stability dependence was confirmed if there were more obstacles upstream as the wind direction incident on the buildings changed. The results highlight the potential for future wake studies using multiple DWLs deploying both vertical and horizontal scan patterns. Dispersion models should incorporate the effect of a complex urban canopy within which tall buildings are embedded.

Modelling the effects of vegetation and urban form on air quality in real urban environments: A systematic review of measurements, methods, and predictions

City-Scale Spatio-Temporal Modeling of 5G Downlink Exposure of Users and Non-Users by Ray-Tracing in a Real Urban Environment

Performance analysis of smart edge computing orchestration algorithms should be done using a realistic urban simulation environment wherein mobile users are accessing their edge services using a readily available 5G network. In this paper, we investigate the influence of using two different 5G simulation frameworks, which are provided as readily available possibilities to model the access network used to deliver edge computing services. The results show that although both frameworks aim to implement the 5G specifications and are deemed suitable choices for simulating a 5G smart city vehicular environment, there can be significant differences in the obtained macro results. The analysis of the simulation results from two identical studies where the only change is the choice of a 5G simulation framework shows that the obtained average end‐to‐end edge service latency as perceived by edge users can differ up to more than 21 times. The choice of 5G simulation framework is also reflected in the overall generated workload for the edge computing orchestration leading to over 25% more migrations when using 5G‐Sim‐V2I/N compared to Simu5G.

Aim: The article explores the application of virtual reality (VR) and the metaverse environment in education aimed at enhancing safety of road traffic and tourism awareness. The research emphasizes the interconnection between tourism and public safety, particularly in unfamiliar urban or natural environments where tourists are exposed to increased risks due to limited knowledge of local traffic rules and geography. The primary goal of the work is to identify and present the potential of immersive virtual technologies as tools for building competencies in safe behaviour for both everyday road users and tourists undertaking mountain expeditions. The paper further aims to demonstrate how small educational or institutional teams can develop effective VR-based solutions with limited resources. Project and methods: The study is based on a dual approach: a literature review of existing VR systems used for safety education in tourism and traffic contexts, and analysis of a practical case of two original VR applications created by the author. These applications - SafeDrive and Rescuer - were developed using the Unity engine and implemented within the Spatial.io metaverse platform. SafeDrive simulates an urban infrastructure environment in the city of Dąbrowa Górnicza to educate citizens on the principles of traffic safety, while Rescuer guides users through the principles of safe mountain tourism using interactive scenarios tailored for different groups, including seniors and emergency responders. The Gartner Hype Cycle model was employed to assess the maturity and future potential of VR technologies in educational applications. Results: The developed applications demonstrated high usability and functionality in diverse educational environments. Testing across various platforms - from VR headsets to mobile browsers - confirmed broad accessibility and effective performance. Immersive interaction with realistic urban and natural environments enabled users to better understand potential threats and develop appropriate behavioural responses. Conclusions: The analysis highlights that VR and metaverse environments offer substantial potential as modern educational tools in the fields of road safety and tourism. The immersive, interactive, and scalable nature of these technologies enables the creation of realistic, safe, and repeatable training conditions that cannot be achieved through traditional educational means. The paper concludes that immersive virtual platforms should become a permanent element of modern strategies for developing social safety awareness and responsible tourism.